[KeepUpdating] Help to understand parameters

[wingedRuslan]

• [ ] 1) graph_measures/calculate_global_measures returns dics of average_clustering, average_shortest_path_length, assortativity, modularity, and efficiency of G , but the scripts/make_figures/plot_network_measures plots

     df = measure_dict
  
     df = df[['a', 'a_rand',
              'M', 'M_rand',
              'E', 'E_rand',
              'C', 'C_rand',
              'L', 'L_rand',
              'sigma', 'sigma_rand']]

  What are these values 'a', 'M', 'E', 'C', 'L', 'sigma'? I suppose 'E' - 'efficiency', 'M' - 'modularity' but others...

• [ ] 2) scripts/make_figures/plot_sagittal_network What are the inputs G_edge, sagittal_pos, axial_pos?

  def plot_sagittal_network(G,
                       G_edge,
                       sagittal_pos,
                       axial_pos,
                       integer_adjust=3,
                       fractional_adjust=2.5,
                       cmap_name='jet',
                       ax=None,
                       figure_name=None):

  Also there is unresolved reference to sort_partition, this function is not defined within the project and is not available from networkx. So I just can't figure out why it is needed to calculate the number of modules

  # Compute the best partition based on the threshold you've specified in cost
  partition = community.best_partition(G)
  
  # Create a sorted list of communitites (modules) according to their average
  # Y coordinate (front to back)
  module_list = sort_partition(partition, axial_pos)
  
  # Display the number of modules
  size = np.float(len(module_list))

[KirstieJane]

Thanks @wingedRuslan!!

This code comes from https://github.com/KirstieJane/NSPN_WhitakerVertes_PNAS2016 which is the code and data to reproduce "Adolescence is associated with genomically patterned consolidation of the hubs of the human brain connectome" (doi: 10.1073/pnas.1601745113).

You can read the application that funded @Islast to start the re-factoring work for this project at https://whitakerlab.github.io/resources/Mozilla-Science-Mini-Grant-June2017.

The supplementary materials for the PNAS paper (https://www.pnas.org/content/pnas/suppl/2016/07/21/1601745113.DCSupplemental/pnas.1601745113.sapp.pdf) contain a lot of useful information. Specifically the "Structural covariance and network analyses" section at the top of page 6 should answer some of your questions. Figure S6 is also quite useful.

• a : assortativity
• C: clustering
• E: efficiency
• L: shortest path
• M: modularity
• sigma: small world measure

This page is also really useful: https://sites.google.com/site/bctnet/measures/list. Some of this code has been translated into python, but networkx will also do quite a bit of the work for us!

Here's a link to the code (that hasn't been translated to the scona module yet) that explains how to get the different layout coordinates: https://github.com/WhitakerLab/scona/blob/2c0f8ec88338757e9bc0608e0002e618efed8148/scona/make_figures.py#L74. This wikipedia page looks ok: https://en.wikipedia.org/wiki/Anatomical_terms_of_neuroanatomy#Planes_and_axes, but this video might be interesting to watch too: https://www.youtube.com/watch?v=f_hxX_xvHQY.

G_edge is a copy of G but probably with a different threshold. If you show all the edges it looks super messy, so in the PNAS paper I used a 2% threshold: https://github.com/KirstieJane/NSPN_WhitakerVertes_PNAS2016/blob/5c9c46caf91768d4cadec2b24078b640f05d3d76/SCRIPTS/make_figures.py#L777

The sort_partition function came from here: https://github.com/KirstieJane/NSPN_CODE/blob/ba8ac4cf42b1aec76030ce67945d95a625107ff0/networkx_functions.py#L16 🙀 🙀 🙀 That one was really tough to search for 😭.

I hope this is useful!

[KirstieJane]

Here's a figure with continuous colours https://github.com/KirstieJane/NSPN_WhitakerVertes_PNAS2016/blob/master/CT_MT_ANALYSES/COMPLETE/FIGS/COVARS_none/Figure4_LowRes.jpg